A comparison of analytic and knowledge-based approaches to closed-shop scheduling

This paper compares three different approaches to scheduling in a closed-shop environment, making the case for a knowledge-based approach. A manufacturing example from the food industry is used as a vehicle for the presentation. The first approach attempts to find an optimal solution using a mixed integer linear programming formulation, but the size of the problem renders this approach impractical. The second approach uses a spreadsheet program to obtain feasible solutions, but imbedded assumptions in the heuristics used allow it to be used only for simple demand patterns. The third approach employs expert systems technology. It includes several heuristics and takes all constraints into consideration. The solution obtained may not be optimal, but computational tests suggest that it is far superior to both spreadsheet and manual approaches.